Journal of Sociolinguistics 13/2, 2009: 249–263
RESEARCH NOTE
Slaves speak pseudo-Toohoku-ben:
The representation of minorities in the
Japanese translation of Gone with the Wind1
Mie Hiramoto
National University of Singapore
INTRODUCTION
This study provides linguistic evidence that the use – and non-use – of Standard
Japanese (SJ) in literary translation indexes social marginality in two societies
(Japan and Southern American), and is based on socio-economic distribution
rather than actual linguistic distribution. The main focus of this study is
the investigation of the intertextuality and the transduction of the speech of
the minority characters (namely, male and female slaves and poor whites)
in the Japanese translation of Gone with the Wind (GWTW). While it is
certain that the minority characters’ use of non-Standard Japanese – which
strongly resembles the stigmatized Toohoku dialect, or Toohoku-ben (TB) – is a
translation of the original non-Standard English (SE), the assignment to them of
something resembling a particular regional Japanese dialect reinforces linguistic
inferiorization of the slaves and poor whites, as well as TB speakers. The use of
this pseudo-dialect is an important element in the linguistic representation of
marginal characters and likewise underscores the salient marginality of TB in
Japanese language ideology.
The original English GWTW text employs eye-dialect orthography to portray
the minority characters’ speech, thus the use of non-SJ in the translation may
be understood as interstitial interdiscursivity. However, assigning speech styles
to characters based on a specific regional dialect, namely TB, requires a different
explanation. Following Ochs’ (1990) model of indexicality, I suggest that the
use of pseudo-TB serves to directly index the characters’ non-standardness, and
therefore to indirectly index the stigmatization associated with TB. Furthermore,
the non-normativity reproduced through this intertextual stigmatization due to
the use of pseudo-TB conclusively serves to stereotype the enslaved men and
C Blackwell Publishing Ltd. 2009
9600 Garsington Road, Oxford OX4 2DQ, UK, and 350 Main Street, Malden MA 02148, USA
250
HIRAMOTO
women slave characters as well as poor whites as minorities by highlighting
their second-class status. In her study of Japanese Women’s Language (JWL),
Inoue investigates the use of JWL in two translated novels and concludes that
translations represent ‘an intertextual relationship that “involves two equivalent
messages in two different codes”’ (Inoue 2003: 318). Her study demonstrates
that JWL is reproduced in translation through the use of language ideology
and that this reproduction is accomplished by erasure of certain attributes
that are presumed to exist in the characters’ original states, a process she calls
transduction.
The current study investigates the speech styles associated with nonnormativity among minority characters labeled as (ex-) slaves and poorwhites in GWTW. If SJ is an unmarked variety, the Japanese language used
by the minority characters in the translated novel is highly marked by
linguistic features that serve to distinguish their speech from the speech of
the Caucasian Southern-aristocrat characters. These distinctions are evident
in a number of linguistic expressions at both morpholexical and phonological
levels. The current study begins from this point, taking up these issues and
developing an analysis which relates to the language ideology of Japanese
dialects. My findings show that neither the slaves nor the poor whites use
SJ, and additionally demonstrate the ways in which their speech is modeled
after, without actually accurately portraying, the stigmatized Japanese dialect
TB.
LINGUISTIC PERFORMANCE IN SCRIPTED SPEECH
Scripted speech, e.g. novel dialogue, movie lines, stand-up comedy, and joke
books, often becomes a site of stereotyping. Sociolinguists and anthropological
linguists have reported on several aspects of the language ideology behind
such stereotypes, especially since Hill’s work on Mock Spanish (e.g. Hill 1993,
1998, 2005). Translating scripted speech that is stereotypical in one language
to another may be problematic since, in many cases, scripted speech may be
idiosyncratic in one speech community and thus the full social impact may
not easily penetrate to another. That is, regarding culturally specific issues
and expressions, a source language and a recipient language are unlikely to
share one-to-one mapping in translation. All things being equal, the closest
possible mapping between dialects in two different languages may be between
standard or standardized dialects. In the original English version of GWTW (and
much other scripted speech of non-standard varieties), the language used by
the minority characters, who are slaves or ex-slaves of African descent and
poor whites, is distinguished from the speech style of the main characters, who
are exclusively Caucasian Southern aristocrats. By observing the convention of
assigning a particular linguistic register, such as a specific regional dialect, to
minority characters, the underlying language ideology is represented in both
the original and the translation of the novel. In regards to the translation,
C Blackwell Publishing Ltd. 2009
SLAVES SPEAK PSEUDO-TOOHOKU-BEN
251
Inoue (2003) states that the main characters speak SJ, which represents
expected gender and social norms, while the minority characters speak with
ungrammatical or unspecified regional forms of Japanese to construct their nonnormative identity. The current study takes as its focus not whether minorities
are speaking ungrammatical Japanese or some unspecified regional dialect
but rather what kind of Japanese dialect is assigned to minority characters,
and what values are attached to that dialect. By investigating in detail the
use of particular linguistic tokens in minority-characters’ speech, I am able
to make a claim regarding which dialect (or its closest approximation) is
being evoked. The results of this investigation demonstrate how a linguistically
under-represented Japanese dialect is used to embody a language stereotype
for minority characters by reference to speech styles of an unrelated
origin.
As pointed out by others, in scripted speech, language stereotyping is often
strongly highlighted. For example, concerning the use of linguistically marked
features to highlight the non-standard speech style of a socially subordinate
group, Ronkin and Karn (1999) report on use of written Ebonics texts in
certain websites, in the form of jokes and parodies, and claim that the
Ebonics-like language found therein does not conform to the grammatical
norms of Ebonics. Ronkin and Karn (1999) argue that scripted texts of this
Mock Ebonics are assigned to some imaginary characters, which supposedly
represent speakers of Ebonics. Furthermore, the imaginary Mock Ebonics
speakers were created by individuals who ignored linguistic features true
to real Ebonics. All in all, they claim that Mock Ebonics is a by-product
of anti-Ebonics ideology, which is represented stereotypically by linguistic
features such as a non-standard writing system. In the case of GWTW, given
that the story is based in the southern United States during the American
Civil War, both majority and minority characters are speaking non-standard
American English to a large extent. However, the non-standardness of the
minority characters’ speech is rendered much more strongly through the
use of non-standard orthography, while the Caucasian Southern aristocratic
characters’ speech is represented in more or less Standard English orthography.
Although Scarlett describes Rhett’s speech as having ‘a light Charleston accent’
on several occasions in the novel, there is hardly any trace of the phonological
characteristics of his proposed dialect marked in writing throughout
the novel. Scarlett’s speech, as well, does not show any characteristic Southerndialect orthography beyond the idiosyncratic level. On the other hand, minority
characters’ speech style is not represented in Standard English orthography,
as seen in the examples of the eye dialect and phonetic spelling below.
Example 1 is an excerpt of dialogue between Scarlett and the minority
character Big Sam. Notice that only Big Sam’s speech is marked with the
non-standard orthography, representing some sort of imagined linguistic
characteristics of Southern black servants which presumably corresponds to his
articulation.
C Blackwell Publishing Ltd. 2009
252
HIRAMOTO
Example 1: Big Sam and Scarlett (Mitchell 1936/2007: 297–398)
Big Sam:
Scarlett:
Big Sam:
No’m, us ain’ runned away. Dey done sont an’ tuck us,
kase we wuz de fo’ bigges’ an’ stronges’ han’s at Tara . . . .
Dey specially sont fer me, kase Ah could sing so good!
Yes’m, Mist’ Frank Kennedy, he come by an’ tuck us.
But why, Big Sam?
Lawd, Miss Scarlett! Ain’ you heerd? Us is ter dig de
ditches fer de w’ite gempmums ter hide in w’en de
Yankees comes.
If spoken, the minority characters’ speech could be imagined to match the
orthography very closely, whereas the main, standard-speaking characters’
dialogue would not represent certain phonological features of the American
Civil War-era Southern-English dialects. In many cases, English texts use nonstandard orthography to represent stigmatized varieties such as Creole languages
and regional dialects, and in this the Japanese language is no different.
According to Ochs (e.g. 1990, 1992, 1996) language signals or indexes
sociocultural information for particular communicative events. Her indexicality
principle is explained through ideas of affective stances (mood, attitude, feeling,
disposition, and degrees of emotional intensity) and epistemic stances (knowledge
or belief) (Ochs 1996: 410). To employ Ochs’ idea, for instance, in the Japanese
translation of GWTW, the majority characters’ use of SJ directly indexes their
upper-class social position or sophistication while indirectly indexing their
normative voice. The minority characters’ use of non-SJ, on the other hand,
serves to highlight their lack of sophistication, and therefore to imply nonnormativity. In the following section, I will discuss language-ideology issues
regarding Japanese dialects that are relevant to the translation of GWTW.
JAPANESE REGIONAL/SOCIAL DIALECTS AND IDEOLOGY
Major linguistic research on Japanese dialectology began after the Meiji
Restoration in 1868, and beginning in the early 1900s, Japanese dialectologists
focused on creating a linguistic atlas for different regional-dialect areas as a
part of the language standardization movement (Yasuda 1999: 196). When
engaging in any analysis involving a Japanese regional dialect, it is important to
remember that the Meiji government targeted specific regional dialects (e.g.
the Toohoku, Okinawa and Kyuushuu dialects) under national languagereformation movements which resulted in their nationwide stigmatization (Sakai
1991: 17). The pressure on these dialects caused by hoogen bokumetsu undoo, or
the dialect abolition movement, was extremely heavy between 1902 and 1960
(Yasuda 1999: 16–19). Even in modern Japan, the stigmatization of TB still
exists among the general public. For example, in studies reporting the survey
results concerning Japanese peoples’ attitudes toward regional dialects, TB is
often associated with negative images such as rusty, slow, or unlively while
C Blackwell Publishing Ltd. 2009
SLAVES SPEAK PSEUDO-TOOHOKU-BEN
253
some other dialects like SJ or Kyooto dialect are evaluated with positive remarks
such as sophisticated, classy, or cosmopolitan (e.g. Kobayashi 1995: 34–46;
Satoo 1995: 20–33; Satoo and Yoneda 1999; Shibata 1958: 90; Shibuya 1995).
Since the Toohoku region is vast and has a rich and varied subregional culture
and history, the term TB does not represent a homogenous linguistic group in
northern Japan. However, as with other major dialects and their subdialects, TB
is useful as an umbrella term that covers a number of other subdialects in the
region. The varieties spoken in the Toohoku region are often derisively referred to
as zuu-zuu ben (zuu-zuu dialect, an onomatopoeic term for the dialect’s distinctive
sound) and repeatedly ridiculed for their unique features (Komori 2000; Sakai
1991). Most Japanese who received a post-World War II education have moved
increasingly towards SJ (e.g. Satoo and Yoneda 1999).
Beyond the realities of the distribution of Japanese dialects, characters
portrayed in scripted speech that belong to certain subgroups are bound by
certain sociolinguistic expectations (e.g. age, gender, socio-economic status,
regional affiliation, etc.) to carry out stereotypical linguistic variations in order
to represent their given roles and traits. Kinsui (2003) suggests the idea of
yakuwarigo ‘role language’ used by imaginary characters in Japanese media
(novels, dramas, animes, translations, etc.). He points out that assignment
of linguistic registers helps audiences identify stereotypical images related to
the roles. In short, he claims that SJ is reserved for protagonists while some
social or regional dialects (or even pseudo dialects) are used by side characters
with specific attributes. In addition, Kinsui (2003) claims that Southern slave
characters in American novels such as GWTW or Uncle Tom’s Cabin are assigned
a country speech style that resembles TB. Stereotypical images and imaginary
characteristics such as ‘normative’ versus ‘non-normative’ are reinforced by use
of specific linguistic registers such as yakuwarigo.
Within Japanese society, SJ is widely understood to be a special register of
linguistic resource reserved for specific personalities. That is, the use of SJ is
socially expected of, and thus assigned to, a limited group of people who are
considered to be stereotypically normative standard-Japanese people. The specific
social register of SJ has traditionally been associated with a group of educated,
non-working-class people and this linguistic ideology endows speakers of this
register with sophisticated qualities. This paper proposes to provide additional
evidence of inequality between upper- and lower-class characters in the Japanese
translation of GWTW, as characterized by linguistic inequality based on the
assignment of standard and non-standard varieties. Further, I claim that what
is described as ‘ungrammatical forms’ or ‘some unspecified regional dialect’
by Inoue (2003) is actually modeled after a specific regional dialect, namely
TB, and that by being made to speak a linguistic variety which resembles
this prototypically stigmatized dialect, the minority speakers in GWTW are
thus made to create their non-normative identity as black servants and poor
whites.
C Blackwell Publishing Ltd. 2009
254
HIRAMOTO
DATA AND RESULTS
A major methodological difference between Inoue’s (2003) work on translation
and transduction of GWTW and the current study lies in the types of data
used for analysis. While Inoue (2003) uses a qualitative (and of course, more
anthropological) analytical method, the current study adds quantitative data
analysis which involves an investigation of linguistic tokens. By doing so, I
try to offer additional evidence that Inoue’s (2003) observation reflects on
not only some of the key slave characters in the Japanese translation but
also on the minority characters in general, including the non-aristocratic
Caucasian characters. Also, I demonstrate the non-standard variety used by the
minority characters is not ungrammatical but rather modeled after a particular
stigmatized regional dialect. For this study, I selected tokens that are easily
distinguishable from SJ forms which appeared relatively frequently in the data
including:
1.
2.
3.
4.
5.
first person pronouns (1PPs);
copula + sentence final particle (SFP);
vowel coalescence;
pre-nasalization; and
a merger of high front vowel (i∼u) after /s/.
As TB is a regional dialect that covers a broad geographical area in the
northern part of the main island of Japan, there are many subgroups of TB
varieties that differ considerably from each other. However, for the purpose of
highlighting convincing evidence that the non-standard speech variety used
by the minority characters is a pseudo-dialect which resembles something
like TB, the investigated features here focus more on the stereotypical pan-TB
characteristics rather than the detailed features of TB subgroups.
The selected linguistic tokens were systematically quantified and coded for
careful data analysis. For the current study, all dialogue by the selected minority
characters (black house servants, black field hands, black coachmen, and poor
whites) was collected from all five volumes of the Japanese translation of the
novel GWTW and the target tokens were coded for analysis. Depending on roles,
the total amount of dialogue per character, or type of character, and thus the
total number of linguistic tokens under investigation, varied greatly. Servants
who were close to the main Caucasian characters had larger roles in the story,
and therefore their linguistic tokens were naturally more numerous compared
to other characters whose roles were not so significant. For example, Scarlett’s
nanny, Mammy, had the largest role of any of the servants; as a result, her
utterances composed 42.5 percent of the data used for the analysis of this study.
Altogether, the data used are inclusive of utterances by five female slaves (55.7%
of the total-data volume), six male slaves (30.7% of the total-data volume), and
two male poor whites (13.5% of the total-data volume).
The first linguistic data to be discussed concern the use of 1PPs. The total
number of 1PPs may appear to be less than expected; however, it is important
C Blackwell Publishing Ltd. 2009
SLAVES SPEAK PSEUDO-TOOHOKU-BEN
255
to bear in mind that Japanese, unlike English, commonly avoids the overt use of
pronouns. Inoue (2003: 326–327) uses examples from two characters, Mammy
and Peter, to exemplify the minority characters’ speech styles, pointing out that
both characters, regardless of their gender differences, employ washi as their 1PP.
Use of washi is predominantly recognized as being reserved for a provincial accent
for elderly male authority figures. However, in some regions (e.g. Hiroshima,
Yamaguchi, Nagano, or Aichi prefectures), washi is used predominantly by both
male and female speakers. According to the data, female characters use washi
about 70 percent of the time and watashi for most of the remaining 30 percent,
while the male characters use washi almost categorically.
While most of the minority characters use washi regardless of their gender,
this form was hardly used by the main characters, in particular the main female
characters, who never used this form. On the other hand, prototypical elderly
male authority figures in the story, such as Scarlett’s father, military officers,
and other older men in respected positions used washi. The use of washi by the
upper-class characters does not sound non-standard, but rather adds dignity
to their speech style within the rules of SJ conversation. On the other hand,
washi as used by the minority characters does not sound dignified because they
are not speaking in the SJ register to begin with. Ora, the stereotypical TB 1PP
(Kanno and Iitoyo 1967/1994; Ooshima 1959/1994), was not used by any of
the minority characters in the translated novel. The stereotypical feminine 1PP,
atashi, was never used by the female minority characters; atashi was, however,
the 1PP most frequently used by Scarlett in the translation. All in all, none of the
female minority characters were assigned the feminine form, atashi, throughout
the text.
The second item to be discussed is the use of sentence-final forms of the
copula/polite verb-ending desu/masu and the SFPs. A distribution of polite
verb endings in the minority-characters’ utterances in the translated novel
includes desu/masu with or without SFPs as well as the non-standard polite
endings gesu/gasu. SFPs da and bee, which are both stereotypical TB forms
(Gaubatz 2007: 140; Iitoyo 1962/1994; Kanno 1982: 386), as well as yo
and ka, which are common standard forms, were observed in the data after
the desu/masu copulas. Although SFPs function as discourse (or phatic) markers
and appear frequently in Japanese, depending on context, it is also natural for
Japanese sentences to end in their bare forms without any SFPs. While there
is a grammatical marking SFP like the ka ‘interrogative marker’, a discourse
marker conveys a speaker’s attitudes or feelings to indicate interpersonal states
rather than grammatical ones. In the data used for this study, different types of
polite verb-ending forms were found: desu/masu + da, masu + bee, desu/masu +
other SFPs, desu/masu bare forms, and gasu/gesu. In comparison to standard
desu/masu verb endings, gasu and gesu are non-standard polite verb-ending
forms commonly associated with rough men or blue-collar workers. As I will
discuss, these endings conform, to varying degrees, to the grammar of either TB
or SJ.
C Blackwell Publishing Ltd. 2009
256
HIRAMOTO
Inoue’s (2003: 327) study provides examples of what she interprets as the
‘ungrammatical’ speech style of some of the main minority characters (Mammy
and Peter):
Example 2: Peter (Mitchell 1977, vol. 4: 116 in Japanese translation; Mitchell
1936: 674 in the original; adopted and modified from Inoue 2003: 327)
Sukaaretto-joosama, arigatoogozeemasu-da (1). Sore wa, washi
is-∗ POLITE-COP-PLAIN
nimo wakatte-iru-da (2).
understand-∗ PLAIN-COP.PLAIN
Yes’m, thankee kinely, Ma’m. Ah knows it . . . .
She explains the ungrammaticality in the above example as follows:
[The minority character is] assigned an ‘ungrammatical’ speech style. It is
highlighted by, for example, a polite verb-ending, gozeemasu, which is considered
non-standard or dialectal for its coalescence of vowel form ai (the standard
form is gozaimasu) to ee (gozeemasu). Furthermore, in standard Japanese, it is
‘ungrammatical’ to attach a plain copula, da(yo), to polite-verb endings such as
masu and desu. (Inoue 2003: 327; emphasis in original)
In Example 2 above, the original English shows an example of ungrammaticality
such as ‘Ah knows it . . . .’ from ‘I know it . . . .’ As Inoue explains, in the
Japanese translation above – as marked with an asterisk (understand-∗ PLAINCOP.PLAIN) – the plain copula da is ungrammatical when following not only
masu/desu verb endings but also plain verb endings. When interpreted as a plain
copula, da in Example 2 makes both verb endings (1) and (2) ungrammatical.
The assignment of ungrammatical forms such as (1) and (2) may be an
interpretation of ungrammaticality in Japanese from of the original. However,
going beyond Inoue’s idea that the ‘da’ in Example 2 is ungrammatical, I offer
another interpretation of da in the example seen above. In TB, there is a SFP da
(which is closer to the standard-form SFP yo); following the polite or plain verbending forms, it expresses a speaker’s feelings such as confirmation, intention,
conclusion, etc. If the two cases of da in Example 2 are treated not as the
plain copula form, but as a SFP, both verb endings (1) and (2) become, not
ungrammatical, but an expression used in TB. Whether da in the example is an
ungrammatical structure or grammatical TB, the Japanese translation highlights
decidedly non-standard elements of the minority characters’ speech. However,
there are other examples that suggest minority speakers are speaking something
like TB.
The data show that the common TB form, desu/masu + da outnumbered
other polite verb-ending forms among the minority slave characters, this ending
comprising a total of 86.6 percent (females) and 69.6 percent (males) of the
total verb-ending tokens. Additionally, use of serial SFPs is a common feature in
conversational Japanese, where they function, in general, as discourse markers,
as in oishii yone ‘tasty, isn’t it?’ as opposed to just oishii ‘tasty’. Among the majority
C Blackwell Publishing Ltd. 2009
SLAVES SPEAK PSEUDO-TOOHOKU-BEN
257
characters, the most frequently used combinations of the polite verb ending and
a single SFP was desu/masu + ne. This result is expected from the data of speakers
of SJ, as ne is the most commonly used discourse particle in modern Japanese,
including standard and many regional varieties. However, this most commonly
used SFP, ne, did not appear alone as an SFP in the minority characters’ data.
That is, whenever minority characters used ne, it followed the TB form da as
in desu/masu + dane. Concerning the SFP ne (following the polite-verb-ending
forms), the minority characters were only assigned the non-standard form, dane,
which can be interpreted as a TB form.
Concerning the use of other verb-ending forms, gasu and gesu are reduced
forms of the formal-polite form gozaimasu via a medial form gansu. In today’s
dialectology, gasu and gesu are predominantly associated with TB (Komatsushiro
1969: 45). The data show that, among the three types of minority characters
(male slaves, female slaves, and poor whites) the male slave characters used this
form the most in the Japanese translation, followed by the poor whites, while
none of the female slave characters used gasu/gesu.
While the data concerning the SFPs is certainly telling, let us examine
additional linguistic forms used by the minority characters. In Example 2
above, pronunciation used by a male slave character, Peter, was introduced;
throughout the novel, he pronounces the standard polite ending form gozaimasu
as gozeemasu. In this case, ai∼ee, is non-standard or dialectal. The ai∼ee
coalescence is, indeed, one of the stereotypical phonological features of TB and
the coalescence is not only limited to the phrase gozaimasu but also occurs with
other vocabulary containing the diphthong ai (Hidaka 2003: 38; Katoo 1969:
122; Kindaichi 1967). For the purpose of this study, the number of occurrences
of the ai∼ee coalescence among the minority speakers was quantified for some
commonly used expressions: the polite verb ending, gozaimasu∼gozeemasu;
the polite imperative marker, kudasai∼kudasee; the negation marker,
-nai∼-nee.
All the minority characters preferred the form ee over ai almost categorically
when using these expressions. Similarly, it would seem highly unnatural for
the main characters to be uttering phrases like gozeemasu or kudasee in the
Southern-aristocratic context since these expressions cannot be taken as a SJ
register. However, it would not be surprising for SJ-speaking males to change
pronunciation of the negation nai to nee in colloquial SJ to mark roughness
or toughness as in shitamachi or beranmee kotoba of Tokyo downtown dialect
(Mitsui 2003: 54; Yoshioka 1995: 48). Thus, it is worth noting that all of the
main male characters used nai as a negation marker and never used nee even in
scenes involving some violence (e.g. fights, riots, war), where nee would be quite
natural.
Another example of TB phonology, prenasalization, was often seen in the
minority speakers’ data. Generally, the prenasalization takes place before the
voiced alveolar stop (d) and the palatal glide (y) in TB (Hanzawa 2003: 42;
Tokugawa and Grootaers 1951). For example, words such as kodomo ‘child’
C Blackwell Publishing Ltd. 2009
258
HIRAMOTO
or koya ‘shack’ are often pronounced [kondomo] or [konya] in many Toohoku
regions. Tokens used for this study were derived from occurrences of a negative
discourse filler iya ‘no, but’, an exclamation oya ‘oh my’, and a negative
conjunction kedo ‘but’. The results show that whenever these expressions were
used, the minority characters had a tendency toward prenasalizing the palatal
glide (y) and voiced alveolar stop (d), at least 70 percent of the time. That is, the
minority characters used inya, onya, and kendo over iya, oya, and kedo. Although
the results discussed up to this point may suggest that the minority speakers
are using TB in the translation of GWTW, the assigned non-standard variety
is more of a pseudo-TB than authentic TB. For example, in authentic TB, one
would expect other voiced stops and palatal glides to be prenasalized in the
appropriate environments. However, except for some idiosyncratic exceptions,
repeated instances of prenasalisations only took place in the three words
mentioned above; thus, overall occurrences of the prenasalized tokens might
have been used just to produce something that sounds superficially like TB.
This point will be further discussed later along with other supporting examples.
Nonetheless, concerning these tokens, the numbers for prenasalization in
the data are in favor of this specific TB phonology among the minority
speakers.
A final example of TB features in the data that I will focus on is a merger of
high front vowels. The high front vowel /i/ merges with the high back vowel /u/
after the voiceless alveolar fricative /s/; thus, a word such as /susi/ [suši] ‘sushi’
in non-TB becomes [susu] in many of the Toohoku regions (Kanno and Iitoyo
1967/1994; Katoo 1969: 118; Kobayashi 2003: 36). There is one instance of
this feature in the data, spoken by one of the male slave characters, Moses, who
pronounces /shingari/ [šingari] ‘last’ as [sungari].
Example 3: Moses (Mitchell 1977, vol. 2: 181 in Japanese translation; Mitchell
2007: 305; parentheses in the original)
Sungari (shingari) no hoosha desuda, Sukaaretto-joosama. Zutto ushiro
no hoodesudayo.
Back wid de las’ cannon, Miss Scarlett. Back dar!
Even though there is only one instance of this token, the merger feature is worth
mentioning due to the fact that this form is one of the most stereotypical features
of TB, known as zuu-zuu-ben.
The findings from the data show that the minority characters used non-SJ
forms much more than the SJ forms. The common characteristics among the
characters were the use of 1PP (the non-SJ washi), vowel coalescence (TB ai∼ee),
and prenasalization (TB inya, onya, kendo). As for the copula and SFP use, female
slave characters predominantly used the polite verb endings desu/masu + TB SPF
da while the male slave characters also used the same TB form as well as the TB
polite verb endings gesu and gasu. Although the poor white characters mainly
used the polite verb endings desu/masu + SJ SPF such as yo or ka, their use of
C Blackwell Publishing Ltd. 2009
SLAVES SPEAK PSEUDO-TOOHOKU-BEN
259
the observed linguistic features remained predominantly non-standard. All in
all, the TB features were used inconsistently by all of the minority characters.
Thus, I conclude that they are assigned a non-standard variety of Japanese
that resembles TB in the translation, whereas they were supposedly speaking
the lower-class Southern dialect in the original. In the Japanese translation of
GWTW, SJ is reserved for the main characters while the minority characters
use some kind of inaka-ben ‘country dialect’ that resembles TB. Therefore, I
claim that the main characters’ normativity in the original GWTW is aligned
with that of Japanese by the assignment of SJ; moreover, the use of pseudoTB by the minority characters highlights not only linguistic but also social
non-normativity.
LINGUISTIC INFERIORIZATION AND INDEXICAL MEANINGS
Siegel’s (1995) notion of linguistic inferiorization is relevant to pseudo-TB usage
here, although the nature of the data differs. Based on his study regarding
two different linguistic varieties in Fiji – one dominant and one subordinate –
Siegel (1995: 104) comments on ‘linguistic inferiorization’ by a dominant group
against a subordinate group. He states that mocking or inferiorization of a certain
linguistic variety ‘arises from ridicule or derision of other people, particularly
those who are considered enemies or those who are thought to be inferior’
(Siegel 1995: 104). By assigning SJ to the main characters and by assigning
something that resembles a widely stigmatized non-standard dialect, namely
pseudo-TB, to the minority characters, linguistic inferiorization is demonstrated
throughout the Japanese translation. Although there are a number of TB features
used in their speech, the register used by the minority characters is not an actual
regional dialect, but rather a linguistic mockery that refers to TB but may just as
easily be interpreted as being ungrammatical. That is, the minority characters
are assigned a linguistic variety that evokes something like a stereotypically
stigmatized Japanese regional dialect, TB, while failing to actually be even
this inferior dialect. This association simultaneously accomplishes linguistic
inferiorization of both the minority characters and actual speakers of TB. As
with other linguistic mockeries, which share a similar lack of concern regarding
correct grammatical construction, pseudo-TB is a linguistic mockery of an
existing dialect which is widely perceived to be an antonym of SJ. As shown
in the data detailed in this paper, realizations of pseudo-TB are most often the
result of a mixing of frequently occurring linguistic tokens of the stereotypical
non-standard features such as SFPs, vowel coalescence, prenasalization, and
merging of high vowels. These dialectal features are combined haphazardly,
even with a non-TB 1PP, washi, which results in the production of something
that sounds superficially similar to TB.
As mentioned above, in the case of washi, when used by elderly-male-authority
figures like Gerald O’Hara (Scarlett’s father), an Army Captain, or Dr. Mead
(the O’Hara family’s physician), it indexes dignity and masculinity. Moreover,
C Blackwell Publishing Ltd. 2009
260
HIRAMOTO
because of the authoritative tone of this 1PP, Kinsui (2003, 2007) refers to
this type of washi as a roojingo ‘elderly speech’ or hakasego ‘scientists’ speech’
in his explanations of role language. At the same time, when washi is used in
non-standard contexts, especially by both females and males, it indexes regional
(non-standard) identity as is a widespread dialectal 1PP in different regions
of Japan (e.g., Hiroshima, Yamaguchi, Nagano, or Aichi prefectures). As used
by both female and male minority characters in the Japanese translation of
GWTW, washi functions as an undignified form because the characters are
not using SJ. Additionally, this highlights that they belong to a lower social
class. However, the same form of 1PP, washi, used by authoritative male figures
functions as a dignified form because the character is speaking SJ and its use in
this case indicates that they belong to an upper social class. This bi-directional
indexicality makes the point that the indexical process of the interpretation
of social meanings of a linguistic form can generate different types of erasure
depending on texual context. For example, by highlighting his dignified position
in the story with the 1PP washi, Scarlett’s father’s own identity as a Southern
aristocrat of Irish heritage undergoes a process of erasure by being aligned with
other Caucasian upper-class male figures who are supposed to share different
identities (e.g. a military officer and a physician). A similar process takes place
among the minority characters and washi here aligns the minority characters
with one another regardless of their gender, skin color, or social background.
Although the same form, washi, helps to categorize and to establish social
identities of the characters, the two different registers – SJ and non-SJ – clearly
index different things when used by characters occupying different positions in
the social hierarchy.
As mentioned earlier, the seeming inequality arising through the pairing of
Southern English in the original GWTW with SJ in the translation cannot be
explained away through simple mapping of one language onto another. The
cultural differences embedded in the two languages give rise to a number of
interdiscursive differences and make such a straightforward correspondence
impossible; such concepts as Southern dialect or slave speech do not have
any correlations in Japanese. What, then, are the reasons for the existing
correlations chosen by the translators? I argue that the characters’ attributes
were aligned ideologically between the original and the translated languages.
The main characters in the translation spoke SJ not because SJ is the Japanese
language equivalent of Southern English, but because the main characters
are wealthy and privileged and were designed to fit the intertextually derived
ideals that language ideology finds to correspond with SJ. In keeping with
this analysis, the unsophisticated, seemingly ungrammatical language of the
slaves and poor whites, the language which highlights their lack of education
and refinement, was linked with TB not because the people of the Toohoku
area were historically slaves, or involved in a bloody civil war as infantry
soldiers, but because the attributes commonly associated with TB are those
of a prototypical inaka-ben ‘country dialect’ (cf. Kinsui 2003) or a variety
C Blackwell Publishing Ltd. 2009
SLAVES SPEAK PSEUDO-TOOHOKU-BEN
261
that evokes an image of someone who is rough, unpolished, and lacking
education.
Hill (2005: 114), referring to Ochs’ idea of indirect indexicality, says that ‘a
feature of the communicative event is evoked indirectly through the indexing
of some other features of the communicative event’ (Ochs 1990: 295). Just
as with Hill’s (2005: 114–115) application of Ochs’ idea to her analysis of
the negative stereotyping of Mock Spanish, pseudo-TB indirectly indexes the
destructive labeling of its speakers, and its ability to do so is due to the overt
‘production and reproduction of deeply naturalized and presupposed elements’
(Hill 2005: 115) of Japanese language ideology.
NOTE
1. My sincere thanks go to the journal’s editors and reviewers for helpful suggestions
and comments on earlier versions of this paper. I am greatly indebted to Laurie
Durand, Benjamin George, Lionel Wee, Timothy Vance, and Andrew Wong for their
invaluable assistance and encouragement at various stages of this project.
REFERENCES
Gaubatz, Thomas Martin. 2007. Translating dialect in Mark Twain’s The Adventures
of Huckleberry Finn. In Satoshi Kinsui (ed.) Yakuwarigo Kenkyuu no Chihei. Tokyo:
Kuroshio Shuppan. 125–158.
Hidaka, Mizuho. 2003. Shojiten furusato no kotoba: Akita-ken. Gekkan Gengo 32:
38–39.
Hill, Jane. 1993. Hasta la vista, baby: Anglo Spanish in the American southwest.
Critique of Anthropology 13: 145–176.
Hill, Jane. 1998. Language, race, and white public space. American Anthropologist
100: 680–689.
Hill, Jane. 2005. Intertextuality as source and evidence for indirect indexical
meanings. Journal of Linguistic Anthropology 15: 113–124.
Iitoyo, Kiichi. 1962/1994. Ishi, suiryoo o arawasu ‘bee’ ni tsuite (1): Fukushima-ken
ni okeru. In Fumio Inoue, Kooichi Shinozaki, Takashi Kobayashi and Takuichiroo
Oonishi (eds.) Toohoku Hoogen-koo vol. 3: Iwate-ken, Miyagi-ken, Fukushima-ken.
Tokyo: Yumani Shoboo. 364–374.
Inoue, Miyako. 2003. Speech without a speaking body: ‘Japanese women’s language’
in translation. Language and Communication 23: 315–330.
Kanno, Hiroshi. 1982. Fukushima-ken no hoogen. In Kiichi Iitoyo, Toshizumi
Hino and Ryooichi Satoo (eds.) Hokkaidoo, Toohoku Chihoo no Hoogen. Tokyo:
Kokushokankookai. 363–397.
Kanno, Hiroshi and Kiichi Iitoyo. 1967/1994. Gengo seikatsu. In Fumio Inoue,
Kooichi Shinozaki, Takashi Kobayashi and Takuichiroo Oonishi (eds.) Toohoku
Hoogen-koo vol. 3: Iwate-ken, Miyagi-ken, Fukushima-ken. Tokyo: Yumani Shoboo.
312–347.
C Blackwell Publishing Ltd. 2009
262
HIRAMOTO
Katoo, Masanobu. 1969. Onin ni tsuite. In Masao Hirai and Munemasa Tokugawa
(eds.) Hoogen Kenkyuu no Subete. Tokyo: Shibundoo. 113–129.
Kindaichi, Haruhiko. 1967. Zenkoku Hoogen Shiryoo vol. 2: Kantoo, Kooshinetsu- hen.
Tokyo: NHK Publications.
Kinsui, Satoshi. 2003. Vaacharu Nihongo: Yakuwarigo no Nazo. Tokyo: Iwanami
Shoten.
Kinsui, Satoshi. 2007. Yakuwarigo Kenkyuu no Chihei. Tokyo: Kuroshio Shuppan.
Kobayashi, Takashi. 1995. Juuminishiki ni miru hoogen shikoo, kyootsuugo shikoo:
Sendai-shi. Gekkan Gengo 24: 34–46.
Kobayashi, Takashi. 2003. Shojiten furusato no kotoba: Miyagi-ken. Gekkan Gengo
32: 36–37.
Komatsushiro, Yuuichi. 1969. Kakuchihoo no hoogen: Iwateken, Nanbu-Hiraizumichoo. In Masao Hirai and Munemasa Tokugawa (eds.) Hoogen Kenkyuu no Subete.
Tokyo: Shibundoo. 45.
Komori, Yooichi. 2000. Koe o kaku. In Konomi Ara and Michiko Tanigawa (eds.)
Kyookai no Gengo: Chikyuuka/Chiikika no Dainamikusu. Tokyo: Shin’yosha. 124–
140.
Mitchell, Margaret. 1936/1977. Gone with the Wind vol. 1. Trans. Yasuo Ookubo and
Michinosuke Takeuchi. Tokyo: Shinchoosha.
Mitchell, Margaret. 1936/2007. Gone with the Wind. New York: Scribner.
Mitsui, Harumi. 2003. Shojiten furusato no kotoba: Tokyo-to. Gekkan Gengo 32: 54–
55.
Ochs, Elinor. 1990. Indexicality and socialization. In James W. Stigler, Richard A.
Shweder and Gilbert Herdt (eds.) Cultural Psychology. Cambridge, U.K.: Cambridge
University Press. 287–308.
Ochs, Elinor. 1992. Indexing gender. In Alessandro Duranti and Charles Goodwin
(eds.) Rethinking Context: Language as an Interactive Phenomenon. Cambridge, U.K:
Cambridge University Press. 335–358.
Ochs, Elinor. 1996. Linguistic resources for socializing humanity. In John Gumperz
and Stephen Levinson (eds.) Rethinking Linguistic Relativity. Cambridge, U.K:
Cambridge University Press. 407–438.
Ooshima, Ichiroo. 1959/1994. Fukushima-ken Hinoemata-mura hoogen no seikaku
ni tsuite. In Fumio Inoue, Kooichi Shinozaki, Takashi Kobayashi and Takuichiroo
Oonishi (eds.) Toohoku Hoogen-koo vol. 3: Iwate-ken, Miyagi-ken, Fukushima-ken.
Tokyo: Yumani Shoboo. 49–166.
Ronkin, Maggie and Helen Karn. 1999. Mock Ebonics: Linguistic racism in parodies
of Ebonics on the Internet. Journal of Sociolinguistics 3: 360–380.
Sakai, Naoki. 1991. Voices of the Past: The Status of Language in Eighteenth-Century
Japanese Discourse. Ithaca, New York/London: Cornell University Press.
Satoo, Kazuyuki. 1995. Hoogen shuryuu shakai, chiiki kooseiin no tayooka to sono
gengoishiki: Hirosaki-shi. Gekkan Gengo 24: 20–33.
Satoo, Kazuyuki and Masato Yoneda. 1999. Doonaru Nihon no Kotoba: Hoogen to
Kyootsuugo no Yukue. Tokyo: Taishuukan.
Shibata, Takeshi. 1958. Nihon no Hoogen. Tokyo: Iwanami Shoten.
Shibuya, Katsumi. 1995. Shinjoo to wakimae ishiki no shoototsusuru tokoro: Kyootoshi. Gekkan Gengo 24: 110–121.
Siegel, Jeff. 1995. How to get a laugh in Fijian: Code-switching and humor. Language
in Society 24: 95–110.
Tokugawa, Munemasa and Willem Grootaers. 1951. Hoogenchirigaku Zushuu. Tokyo:
Akiyama Shoten.
C Blackwell Publishing Ltd. 2009
SLAVES SPEAK PSEUDO-TOOHOKU-BEN
263
Yasuda, Toshiaki. 1999. Kokugo to Hoogen no Aida: Gengokoochiku no Seijigaku. Kyoto,
Japan: Jinbun Shoin.
Yoshioka, Yasuo. 1995. Shuto hoogen no ikoo to gengoishiki, gengohenka: Tokyo.
Gekkan Gengo 24: 48–60.
Address correspondence to:
Mie Hiramoto
Department of English Language & Literature
Faculty of Arts & Social Sciences
National University of Singapore
Blk AS5, 7 Arts Link
Singapore 117570
[email protected]
C Blackwell Publishing Ltd. 2009